Aims/objectives: To compare, using two pH-cycling models, the de/remineralisation effects of children's toothpastes on primary teeth. Design: In vitro single-section and pH-cycling models. Methods: Primary teeth were placed in demineralising solution for 96 hours to produce artificial carious lesions 60–100μm deep. They were cut into 100μm thick sections and assigned to 6 groups. Sections in Groups A and D were exposed to a non-fluoridated toothpaste, those in Groups B and E to half-pea-sized (0.16g) and those in Groups C and F to pea-sized portions (0.32g) of a 500ppm F toothpaste. pH-cycling Model I (Groups A, B, C), without added fluoride, ran for 7 days, while Model II (Groups D, E, F), with 0.25ppm F, ran for 10 days. Outcome measurements: Lesions were evaluated using polarised light microscopy and microradiography. Results: Lesions in Groups B and E progressed by 64% and 61 %, respectively, while those in Groups C and F progressed by only 19% and 23% respectively. Conclusions: Both 10-day and 7-day pH-cycling models were suitable for studying carious lesion progression in primary teeth (the demineralising and remineralising solutions of the 10-day cycling model contained 0.25ppm F). A pea-sized portion (0.32g) of 500ppm F toothpaste slowed down the demineralisation progression better than a half-pea-sized portion.